Table 5. Associations between real election outcomes and face-based inferences of corruption.

	Average Agreement		Cross-country Comparison
	U.S. Electiona	Korean Electionb	U.S—Koreanc
All participants (N = 80)	49.18%	52.21%	-3.03%
SD	9.47%	8.43%	t (79) = -2.07
95% CI	[47.42%, Inf)	[50.64%, Inf)	[-5.94%, -0.11%]
Caucasian participants (N = 40)	47.50%	52.46%	-4.96%
SD	10.28%	8.72%	t (39) = -2.29
95% CI	[44.76%, Inf)	[50.14%, Inf)	[-9.33%, -0.59%]
Korean participants (N = 40)	50.86%	51.95%	-1.09%
SD	8.38%	8.24%	t (39) = -0.56
95% CI	[48.62%, Inf)	[49.76%, Inf)	[-5.05%, 2.86%]

^aAverage agreement between U.S. election outcomes and face-based inferences of corruption, and its one-sided t-test against chance level 50%.

^bAverage agreement between Korean election outcomes and face-based inferences of corruption, and its one-sided t-test against chance level 50%.

^cTwo-sided t-tests on the average agreement across U.S. and Korean elections.